Anthropic IPO talks: ~$2T ask with Nvidia as possible anchor ↗
Reported IPO ask (~$2T) is not the Series H private mark ($965B). Separating private primary from IPO ask is the overvaluation horseman’s homework.
PERSISTENT HISTORICAL RECORD · NOT A PROPHECY
The most meaningful articles and signs from 2026 forward, organized by the Four Horsemen— plus foundational Owen A. Lamont / Acadian Owenomics pieces that inspired the framing. Curated, linked, and labeled—built so weekly digests can append without reinventing the archive.
I · OVERVALUATION
Are prices unreasonable versus history and sober expert opinion?
Reported IPO ask (~$2T) is not the Series H private mark ($965B). Separating private primary from IPO ask is the overvaluation horseman’s homework.
Trailing fundamentals versus trillion-scale pricing make the overvaluation horseman concrete even before OpenAI trades.
Primary company disclosure pricing an astonishing future into a private mark—overvaluation evidence without claiming the multiple is ‘wrong.’
II · BUBBLE BELIEFS
Does everyone know it’s expensive—and buy because it may get pricier?
Before: II was editorial judgment only. After: II = 35% Yale individual-investor survey (Lamont-style) + 40% weighted bubble headlines (top outlets or speakers ×5, mid ×2, rest ×1) + 25% editorial (capped at 3 points a week and 8 over a quarter; doubt lowers it only when money is leaving). Oct 31 restated: II 50 → 72; How bubbly 232 → 254/400. The same rescoring was applied to every week from Mar 7 to Oct 31.
Bloomberg (via Business Standard, Oct 30) reports record 5-year CDS closes on Oct 29 for Broadcom (175.4bp) and SpaceX (266.5bp), with Alphabet about 84bp (ICE Data Services). On the /ai-cds board, the eight names with same-vendor readings (Investing.com and S&P Global) show a median 22.4% widening from Sep 29–30 to Oct 26–29. Not every name is at an extreme: SoftBank was 432bp on Oct 20 (Bloomberg via Yahoo Finance, vendor not named) and Nvidia 127bp on Oct 26, below its 130bp high of Oct 23 (both Nvidia figures S&P Global). Lenders charging more to hold AI risk is a skeptic signal, the opposite of “too high but going higher.”
10y from below ~4.8% (~2 weeks prior) to above ~5.17% (highest since ~Jul 2007); fastest one-day rise since Apr 2025. John Roque (22V Research) marks 16 rapid 10y advances since 1970 and claims each preceded some financial disruption (scale varies: SVB ’23 ↔ 1987). Watches: regional banks (KRE), utilities, homebuilders; traders also cite private credit and debt-funded AI datacenter CapEx. Velocity vs level is the point—rate-stress narrative on the mainstream tape.
APPENDIX · BUBBLE BURSTER RECONSTRUCTION
Roque’s full 16 is in a 22V client note and is not public in CNBC. The table below is a Bubble Burster reconstruction from a circulating chart attributed to 22V Research (edited for clarity) plus CNBC’s confirmed examples (1987, SVB 2023). It is not a verified facsimile of the original note — inferred candidates only, not Roque’s official proprietary list.
Chart frame starts ~1970 (post–Bretton Woods / “five decades”); FRED DGS10 goes back to 1962, so the cutoff is discretionary.
| # | Window | What broke (or nearly) | Confidence |
|---|---|---|---|
| 1 | Oct-74 | Franklin National Bank failure | MEDIUM |
| 2 | Feb/Mar 1980 | First Pennsylvania rescue | MEDIUM |
| 3 | Sep-81 | S&P bear market | MEDIUM |
| 4 | Jul-82 | Penn Square Bank failure (contagion vector) | MEDIUM |
| 5 | May-84 | Continental Illinois | MEDIUM |
| 6 | Sep-87 | 1987 crash (Black Monday) · CNBC-named | HIGH |
| 7 | Apr-90 | Nikkei bust / S&L crisis | MEDIUM |
| 8 | Nov-94 | Mexico peso crisis / Orange County | MED–HIGH |
| 9 | Aug-96 | SE Asia (pre–1997 stress) — Timing early vs Jul-97 Asia crisis peak — stretch on date. | MEDIUM |
| 10 | Jan-00 | NASDAQ / tech bust | MED–HIGH |
| 11 | May-06 | Housing / subprime setup — Setup phase; acute GFC breaks came later. | MEDIUM |
| 12 | Dec 2009–Mar 2010 | GFC aftermath / PIIGS sovereign stress | MED–LOW |
| 13 | Dec-13 | EM / taper-tantrum aftermath | MEDIUM |
| 14 | Oct/Nov 2018 | QT / balance-sheet runoff stress | MED–LOW |
| 15 | 2022 | “Everything” multi-asset drawdown (incl. UK LDI) | MED–HIGH |
| 16 | 2023 | SVB / Signature / other regional banks · CNBC-named | HIGH |
16 inferred candidates · not a verified facsimile of the 22V note
Same Owenomics piece dual-tagged under Bubble beliefs: issuance is not just supply—eager buyers who know it is expensive are the greater-fool half of the trade.
Foundational update: Lamont flags epic return dispersion as a bubble-adjacent whirlwind (not itself one of the Four Horsemen). Also covered in Fortune the same week—primary link is Owenomics.
High-credibility dissent enters the mainstream tape—bubble-language and crash framing from a watched skeptic, not anonymous social media.
Primary exchange data: call purchases >52% of retail opening activity in ten mega-cap tech names—buy-anyway behavior you can measure.
Foundational application of the same Four Horsemen to the AI boom: expensive market + bubble beliefs + inflows, but no IPO/issuance wave yet—so boom, not bubble, in Lamont’s read.
Foundational deep-dive on Horseman II: Yale survey evidence that many investors call the market overvalued while still expecting further gains—the greater-fool signature this ledger watches.
Foundational inspiration for this site’s dials: Owen A. Lamont defines the Four Horsemen as overvaluation, bubble beliefs, equity issuance, and investor inflows—the checklist Bubble Burster builds on.
III · ISSUANCE
Are companies and insiders supplying shares into eager demand?
The working ‘final boss’ tripwire slips: filing is not a trading date, and the CEO now explicitly pushes a 2026 listing off the table pending safety work.
A major global exchange listing with a ~$46B top-end valuation and a concrete Sep 17–21 subscription window extends the 2026 issuance procession beyond U.S. AI/space names.
Another outsized 2026 IPO timetable (~$1.6B raise / ~$49B valuation) and explicit retail-democratization framing—issuance appetite outside the AI hectocorns.
Foundational Issuance read: Owenomics on IPOs of doom—how equity supply into eager demand marks a bubble regime. Dual-tagged interest for Bubble beliefs (greater-fool buyers meeting the float).
Dress-rehearsal mega-IPO: enormous primary supply absorbed in a single debut—issuance as market regime, not one-off company story.
Primary paperwork for the thesis tripwire. Filing starts the watchlist clock; it does not start the exit clock until a credible trading date appears.
Precursor event in the AI IPO procession—concrete SEC paperwork ahead of OpenAI’s own filing announcement.
IV · INFLOWS
Are new participants stampeding in—and is leverage expanding?
Retail-leverage / new-entrant froth: FINRA debit balances in customers’ securities margin accounts were $1,453,832M in Aug 2026 (+37.2% YoY vs $1,059,723M), after an all-time record $1,502,072M in Jun 2026 (+49% YoY); Jul $1,417,225M. Robinhood’s Margin Book hit a record $21.6B at Q2 2026 end (+127% YoY; ~4.3x the ~$5B of Q2 2024)—leverage expanding fastest at the retail edge.
Seasonal/cycle context already on the home timeline. Useful calendar awareness—not proof that any horseman is riding.
A high-profile AI bull fund exits public equities after a leveraged drawdown—positioning and forced-sale evidence, not a refutation of the tech thesis.
Forced liquidation in a leveraged AI/chip trade is the kind of unwind headline the framework watches—participation funded by margin, not just conviction.